Biochemistry 1996, 35, 8143-8148
Activation of Ribulose-1,5-bisphosphate Carboxylase/Oxygenase (Rubisco) Involves Rubisco Activase Trp16 Frank J. van de Loo and Michael E. Salvucci* United States Department of Agriculture, Agricultural Research SerVice, Western Cotton Research Laboratory, 4135 East Broadway Road, Phoenix, Arizona 85040-8830 ReceiVed February 28, 1996; ReVised Manuscript ReceiVed April 26, 1996X
ABSTRACT: The role of the N-terminal region of tobacco Rubisco activase in ATP hydrolysis and ribulose1,5-bisphosphate carboxylase/oxygenase (Rubisco) activation was examined by construction of mutant proteins. Deletion of the first 50 amino acids of Rubisco activase almost completely eliminated the ability to activate Rubisco, without changing the ATP-hydrolyzing and self-associating properties of the enzyme. Thus, the N-terminus of Rubisco activase is distinct from the ATP-hydrolyzing domain and is required for Rubisco activation. Directed mutagenesis of the species-invariant tryptophan residue at position 16 inhibited Rubisco activation but not the binding or hydrolysis of ATP. The ability to activate Rubisco was less severely inhibited when Trp was replaced by a Tyr or Phe than by an Ala or Cys, indicating that an aromatic residue at position 16 and particularly a Trp is required for proper activation of Rubisco. Fluorescence quenching of the 7-nitrobenz-2-oxa-1,3-diazole-modified W16C mutant upon addition of nucleotide suggested that position 16 becomes more solvent accessible in response to nucleotide binding. However, changes in the intrinsic fluorescence of truncated and Trp16 mutants upon addition of ATP were similar to those of the wild type, evidence that Trp16 is not the residue reporting the conformational change that accompanies subunit association.
The primary enzyme of photosynthetic carbon fixation, ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco,1 EC 184.108.40.206), is regulated by Rubisco activase, a chloroplastic enzyme that facilitates the release of tight-binding sugar phosphates from Rubisco. Without Rubisco activase, tightbinding sugar phosphates act as inhibitors of Rubisco by either hindering carbamylation of the activator lysine or blocking the substrate, ribulose 1,5-bisphosphate, from binding to the carbamylated enzyme [for reviews, see Portis (1990) and Hartman and Harpel (1994)]. The hydrolysis of ATP by Rubisco activase is required to convert Rubisco from the tight- to the loose-binding form. ATP hydrolysis presumably supplies a source of energy for the conformational changes that reduce the binding affinity of Rubisco for sugar phosphates. However, the nature of the interaction between Rubisco and Rubisco activase is unknown. In previous studies, photoaffinity analogs of ATP were used to identify the subdomains of the ATP binding site of Rubisco activase (Salvucci et al., 1993, 1994). Photoaffinity labeling of the purified tobacco leaf enzyme with ATPγbenzophenone identified peptides in the ATP γ-phosphate binding domain (Salvucci et al., 1993). Breakdown products of 38 and 39 kDa, present in the preparation, also photolabeled and had saturation and protection kinetics identical to those of the full-length 42 kDa polypeptide. The 38 and 39 kDa polypeptides commenced at Ser56 and Phe41, respec* Address correspondence to this author at USDA-ARS, Western Cotton Research Laboratory, 4135 East Broadway Road, Phoenix, AZ 85040-8830. Phone: 602-379-3524 ext. 227. Fax: 602-379-4509. E-mail: [email protected]
X Abstract published in AdVance ACS Abstracts, June 1, 1996. 1 Abbreviations: Rubisco, ribulose-1,5-bisphosphate carboxylase/ oxygenase; NBD, 7-nitrobenz-2-oxa-1,3-diazole; N3ATP, azidoadenosine 5′-triphosphate; ATPγS, adenosine 5′-(γ-thio)triphosphate.
tively, indicating that these breakdown products were truncated at the N-terminus. Thus, the N-terminus of Rubisco activase through Gln55 does not appear to be necessary for ATP binding. Photoaffinity labeling of recombinant tobacco Rubisco activase was conducted with 2- and 8-N3ATP to identify the subdomain involved in binding the adenine base. 2-N3ATP labeled the peptide Asn68 to Asp74, and incorporation into this peptide was prevented by ADP and ATP (Salvucci et al., 1994). The azido-substituted ATP analogs also labeled a peptide near the N-terminus of Rubisco activase. Labeling of this peptide was not prevented by ADP and ATP but was blocked by tryptophan and other indoles. On the basis of these observations, the binding of azidoadenine nucleotides to this site was ascribed to base-stacking interactions with Trp16, a species-invariant tryptophan that is one of the most highly conserved residues in the labeled peptide (Salvucci et al., 1994). Interestingly, modification in the vicinity of Trp16 inhibited Rubisco activation without affecting ATP hydrolysis. A similar effect on the two activities of Rubisco activase was recently reported for a deletion mutant of the spinach enzyme that lacked the N-terminal 12 residues through the analogous Trp residue (Esau et al., 1996). In this study, we examine the involvement of the Nterminus of Rubisco activase, and particularly Trp16, in the interaction between Rubisco activase and Rubisco. For these studies, we use in Vitro mutagenesis to construct a truncation mutant (Trunc) that begins at His51 and mutants with various substitutions at Trp16. Study of the recombinant mutant proteins showed that both types of mutations impair Rubisco activation, without affecting the binding or hydrolysis of ATP. Thus, the N-terminus of Rubisco activase and specifically Trp16 is not necessary for ATPase activity but is required for Rubisco activation.
S0006-2960(96)00490-4 This article not subject to U.S. Copyright. Published 1996 by the American Chemical Society
8144 Biochemistry, Vol. 35, No. 25, 1996 MATERIALS AND METHODS In Vitro Mutagenesis and Protein Purification. A cDNA clone was previously engineered for expression of a mature form (i.e. minus chloroplast transit peptide) of tobacco Rubisco activase (Salvucci & Klein, 1994). This “wild type” clone is similar to the partial clone pJQ11 (GenBank Z14979) sequenced by Qian and Rodermel (1993) and to the fulllength clone NtRCA342 (GenBank U35111) sequenced by G. W. Snyder (unpublished). However, the first five residues of our mature protein sequence (EEKDA) following the initiating M were engineered to correspond to protein sequence data (Wang et al., 1992) and differ from the predicted translation product of NtRCA342 (EQIDV). An N-terminally truncated protein (Trunc mutant) was engineered by PCR, using the forward primer 5′TACTCCCATGGCTGTGTTGCAATCCTACGAATAC3′ and reverse primer 5′CTACTCCCACAAAAGTGCCCTTAAAA3′. The PCR product was digested with NcoI and SalI and cloned into the expression vector pET21d (Novagen2 ). For sitedirected mutagenesis of the Trp16 codon, an NcoI-HindIII fragment coding for the N-terminus of Rubisco activase was cloned into the M13 phage vector M13BM21 (BoehringerMannheim) for oligonucleotide-directed mutagenesis (Amersham). Initially, the W16A mutant was made using the mutagenic oligonucleotide 5′CGACAGTGACAGAGCGAAGGGTCTTGT3′, and an XbaI-HindIII fragment encompassing the mutation was used to replace an XbaI-HindIII fragment of pET21d containing the wild type cDNA cloned at NcoI and SalI sites. Subsequently, a primer was designed for introduction of the W16Y, W16S, and W16F mutations. This primer was 5′CGACAGTGACAGAT(A/C/T)TAAGGGTCTTGTCC3′. A separate primer was designed for isolation of the W16C mutant, 5′CGACAGTGACAGATGTAAGGGTCTTGTCC3′. The W16Y/C/F mutant NcoISmaI fragments were cloned by replacement of the wild type fragment in the pET21d vector. The mutagenized DNA fragments for the Trunc, W16A, W16Y, W16C, and W16F mutants were sequenced to ascertain that no extraneous mutations were introduced during PCR or cloning in the M13 phage. Wild type and mutant Rubisco activase proteins were expressed in the Escherichia coli strain BL21(DE3)pLysS (Novagen) by induction of an (OD600 ) 0.6) M9ZB culture (Studier et al., 1990) with 0.4 mM isopropyl β-thiogalactoside at 37 °C. Cells were harvested 3.5 h after induction. Resultant poor expression (80% of total protein, >50% of soluble protein) by using the host strain lacking the plasmid pLysS. Similar expression was achieved in the pLysS-containing strain by using the vector pET23d, which differs from pET21d by the absence of the lacI gene encoding the lac repressor, and absence of the lac operator from the promoter. Rubisco activase was purified from the pLysS-containing cells essentially as described (Salvucci & Klein, 1994), except that cells were lysed by thawing and then sonicated briefly to reduce the viscosity. 2 Mention of a trademark, proprietary product, or vendor does not constitute a guarantee or warranty of the product by the U.S. Department of Agriculture and does not imply its approval to the exclusion of other products or vendors that may also be suitable.
van de Loo and Salvucci Photoaffinity Labeling. Photoaffinity labeling of recombinant wild type and mutant Rubisco activase with [γ-32P]2-N3ATP was performed in 25 µL reaction mixtures as described previously (Salvucci et al., 1994) using the concentrations of [γ-32P]-2-N3ATP and Rubisco activase indicated in the figure legends. Photolabeled proteins were analyzed by direct autoradiography following separation of the labeled polypeptides by SDS-PAGE (Salvucci et al., 1993, 1994). Assay of Rubisco ActiVase ActiVities. ATPase and Rubisco activation was measured in the presence of 5% (w/v) PEG3350 as described previously (Salvucci, 1992) except that the concentrations of MgCl2, dithiothreitol, pyruvate kinase, and lactate dehydrogenase in the ATPase assay were changed to 5 mM, 5 mM, 4.2 u mL-1, and 6.0 u mL-1, respectively. The concentrations of proteins used in the assays are indicated in the figure legend. Assays were conducted in triplicate (activation) or duplicate (ATPase), and the results presented are the means of representative experiments. Chemical Modification of W16C. Wild type Rubisco activase and the W16C mutant were desalted in 50 mM potassium phosphate (pH 7.5) after precipitation with ammonium sulfate (Salvucci, 1992). The desalted proteins were incubated in the dark at 2 mg mL-1 (47.6 µM protomer) with a 2-fold molar excess of N,N′-dimethyl-N-(iodoacetyl)N′-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylenediamine [(iodoacetyl)-NBD]. After 5 h at 23 °C, Rubisco activase protein was precipitated by addition of ammonium sulfate to 50% (w/v), collected by centrifugation, and then resuspended and desalted in 50 mM Hepes-KOH (pH 7.2), 50 mM sucrose, 2.5 mM MgCl2, and 10 mM EDTA. Approximately 0.8 mol of NBD was incorporated per mole of W16C subunit on the basis of the A500 of the NBD adduct, compared to 0.59 mol/ mol for the wild type. Automated Edman degradation analysis of reverse-phase HPLC-purified, Staphlococcus aureus V8 protease-generated peptides from (iodoacetyl)NBD-modified W16C (Salvucci et al., 1994) verified that approximately 78% of the incorporated NBD was associated with the N-terminal peptide, i.e. the V8 peptide containing Cys16. Presumably, the remaining 22% of the incorporated NBD (i.e. 0.17 mol/mol of subunit) was associated with the four naturally occurring Cys residues. Miscellaneous Techniques. Intrinsic fluorescence was measured using the conditions described in Wang et al. (1993). Fluorescence of the W16C-NBD was measured at 23 °C as described in the figure legend. Rubisco activase protein was determined by the method of Bradford (1976) using bovine serum albumin as a standard. RESULTS In Vitro Mutagenesis and Protein Purification. A cDNA clone engineered to resemble the mature tobacco Rubisco activase protein was expressed in E. coli cells as the wild type Rubisco activase (Salvucci & Klein, 1994). An N-terminally truncated protein (Trunc) was expressed from a derived clone, in which the codon for His51 was replaced with the initiator Met codon. Site-directed mutagenesis was used to replace Trp16 of the wild type protein with Ala (W16A), Tyr (W16Y), Phe (W16F), or Cys (W16C). Wild type and mutant proteins were purified from E. coli cells expressing the appropriate clone by ammonium sulfate precipitation, rate-zonal centrifugation, and anion exchange
Activase Trp 16 in Rubisco Activation
FIGURE 1: Photoaffinity labeling of Trunc and wild type (WT) Rubisco activase with 10 µM [γ-32P]-2-N3ATP. Wild type (10 µg, 9.3 µΜ protomer) and Trunc mutant (10 µg, 10.8 µΜ protomer) protein were photoaffinity labeled in a total volume of 25 µL: (A) autoradiogram and (B) Coomassie-stained gel. The positions of 42 and 44 kDa bands described in the text are indicated.
chromatography. Similar behavior through these various modes of separation indicated that none of the mutations caused major changes in the physical properties of the native enzymes. ATP Binding. When wild type Rubisco activase is photoaffinity labeled with [γ-32P]-2-N3ATP, approximately 2 mol is incorporated per mole of subunit (Salvucci et al., 1994). 2-N3ATP labels two distinct sites on the enzyme, the ATP-hydrolysis site and a separate site near the Nterminus. Modification of the ATP-hydrolysis site has no effect on the electrophoretic mobility of the polypeptide and can be prevented by inclusion of ATP or ADP. In the present study, 32P was covalently incorporated into the 38 kDa polypeptide when the Trunc mutant was photoaffinity labeled with [γ-32P]-2-N3ATP (Figure 1). The modified polypeptide migrated as a single band on SDS-PAGE indistinguishable from the unmodified protein. The Trunc mutant was also modified when photoaffinity labeled with 10 µM [γ-32P]-8N3ATP and was protected against photolabeling by ADP (data not shown). Previous results showed that, at this concentration, 8-N3ATP preferentially labels the ATPhydrolysis site (Salvucci et al., 1994). Thus, the results are consistent with the presence of a functional ATP-hydrolysis site in the Trunc mutant protein and the absence of the N-terminal site. Judging from the intensity of the image on the autoradiogram, mutation of Trp16 to Ala, Tyr, Phe, or Cys had no effect on photoaffinity labeling of the ATP-hydrolysis site of Rubisco activase (42 kDa band) as compared to wild type (Figure 2). However, the W16A, W16Y, and W16F mutations abolished labeling of the N-terminal site (44 kDa band, Figure 2A-D). Slight labeling of this site occurred in the W16C mutant when a higher concentration of photoaffinity
Biochemistry, Vol. 35, No. 25, 1996 8145
FIGURE 2: Photoaffinity labeling of wild type (WT) and Trp16 mutants of Rubisco activase with [γ-32P]-2-N3ATP. Wild type and mutant proteins (7 µg, 6.7 µΜ protomer) were photoaffinity labeled with 50 µM [γ-32P]-2-N3ATP (A and B) or 10 µM [γ-32P]-2-N3ATP (C-H) with no addition (C and D) or with 200 µM ADP (E and F) or 100 µM tryptophan (G and H): (A, C, E, and G) autoradiographs and (B, D, F, and H) corresponding Coomassiestained gels. The positions of 42 and 44 kDa bands described in the text are indicated. Table 1: Kinetic Parameters for ATP Hydrolysis by Wild Type and Mutant Recombinant Rubisco Activasea enzyme
wild type Trunc W16A W16Y W16F W16C
1.07 0.95 1.14 0.95 1.05 0.95
0.124 0.085 0.116 0.116 0.111 0.111
1.52 1.99 1.58 1.53 1.63 1.82
a ATP hydrolysis was assayed spectrophotometrically by following coupled NADH oxidation. All reaction mixtures contained 0.466 µM activase. b (mol of ATP) (mol of activase)-1 s-1. c Half-maximal [ATP] in millimolar.
probe was used (Figure 2A,B). In the presence of 10 µM [γ-32P]-2-N3ATP, labeling of the ATP hydrolysis site (42 kDa band) was protected by 200 µM ADP, for both wild type and Trp16 mutant proteins (Figure 2E,F). In contrast, 100 µM tryptophan provided no protection against labeling of this site in wild type and the Trp16 mutants but protected against labeling of the N-terminal site in the wild type (44 kDa band, Figure 2G,H). These results demonstrate that mutation of Trp16 does not affect binding of an ATP analog to the ATP hydrolysis site but disrupts photoaffinity labeling of the N-terminal site. ATP Hydrolysis. Purified Rubisco activase hydrolyzes ATP, even in the absence of Rubisco. Consistent with unimpaired ATP binding by the Trunc and Trp16 mutant proteins was the observation that these mutants also hydrolyze ATP. The Vmax values, Hill coefficients, and halfmaximal ATP concentrations for ATP hydrolysis by the mutants were similar to those of wild type enzyme (Table 1), indicating that removal of the N-terminus of Rubisco activase, or replacement of Trp16 with Ala, Tyr, Phe, or Cys, had no effect either on the affinity for ATP or on the maximal rate of ATP hydrolysis.
8146 Biochemistry, Vol. 35, No. 25, 1996
van de Loo and Salvucci
Table 2: Initial Rates of Rubisco Activation by Wild Type and Mutant Rubisco Activasea experiment 1 b
wild type Trunc W16A W16Y W16F W16C
9.41 0.33 0.83
% of WT
experiment 2c rate
% of WT
5.81 3.5 8.8
0.58 3.10 3.94 0.62
10.0 53.3 67.8 10.7
Decarbamylated Rubisco complexed with RuBP was incubated at 4 (experiment 1) or 1.5 (experiment 2) mg mL-1 with 0.1 (experiment 1) or 0.2 (experiment 2) mg mL-1 Rubisco activase for 20 s. Rates are calculated by subtraction of the no-activase control. b (mol of Rubisco sites activated) (mol of activase)-1 min-1. c Experiment 2 corresponds to the time course in Figure 3.
FIGURE 3: Rubisco activation by Trp16 mutants and wild type Rubisco activase. Decarbamylated Rubisco complexed with RuBP was incubated at 1.5 mg mL-1 in the absence (2) or presence of 0.2 mg mL-1 wild type (O) or mutant W16F (9), W16Y (0), W16C (4), or W16A (b) Rubisco activase. At the indicated times, a 15 µL aliquot was removed for determination of Rubisco activity.
Rubisco ActiVation. Since selective modification of the N-terminal site of Rubisco activase with 2-N3AMP inhibits Rubisco activation (Salvucci et al., 1994), we tested the Trp16 mutants for the ability to activate Rubisco. The initial rates of activation for the Trunc and W16A mutants were only 3 and 9% of that of the wild type, respectively (Table 2). In a separate set of experiments, the initial rates of activation for the W16A, W16C, W16Y, and W16F mutants were only 10, 11, 53, and 68% of that of wild type, respectively (Table 2). The decreased rates of Rubisco activation in the Trp16 mutants were not caused by a lower affinity for Rubisco since the Km(Rubisco) values of the mutants were only slightly higher than those of the wild type (data not shown). The time course of Rubisco activation by the Trp16 mutants indicated that the W16C and W16A mutations had a marked effect on the final extent of Rubisco activation, whereas the W16Y and W16F mutations had a much smaller effect (Figure 3). ActiVase-ActiVase Interactions. Rubisco activase is a self-associating protein, and its specific activity increases with the degree of aggregation (Salvucci, 1992). To investigate the ability of the W16A and Trunc mutants to undergo self-association, we measured ATPase activity at various concentrations of Rubisco activase. The specific ATPase activity of the mutants increased with the concentration of the enzyme in a relationship similar to that of wild type (data not shown). For example, specific activities increased approximately 50% with an increase in protein concentration from 0.116 to 0.466 µM. ATPase activity of the Trunc and W16A mutant enzymes was additive with wild type enzyme in mixing experiments (data not shown), indicating that interactions between mutant
FIGURE 4: Effect of nucleotide and denaturants on the fluorescence emission of wild type and W16C Rubisco activases derivatized with (iodoacetyl)-NBD. Fluorescence of the derivatized proteins (40 µg) was measured in 100 mM Tricine-NaOH (pH 8.0), 10 mM MgCl2, and 10 mM NaHCO3 in the absence (a, b, d, and e) and presence (c) of 3 M guanidine hydrochloride (Gu-HCl) in a total volume of 1.5 mL. NBD fluorescence was excited at 467 nm. (A) Time course of the fluorescence emission at 537 nm for wild type (d and e) and mutant (a-c) Rubisco activase. ADP (0.2 mM) was added at the time indicated by the arrowhead. (B) The emission spectra of panel A before (a, c, and d) and after (b, c, and e) the addition of ADP.
and wild type enzymes were also normal. When mutant and wild type enzymes were mixed, the mutant enzymes had little effect on Rubisco activation by the wild type enzyme (data not shown). These results demonstrated that mixed complexes of wild type and W16A mutant maintained the ability to activate Rubisco, unlike complexes of wild type and the K247R mutant, a mutant with very low ATPase activity. The K247R mutant markedly inhibited activation of Rubisco by wild type Rubisco activase (Salvucci & Klein, 1994). Changes in Fluorescence of Chemically Modified W16C. To determine if changes in conformation occur in the vicinity of Trp16, the W16C mutant was chemically modified at Cys16 with (iodoacetyl)-NBD, an environmentally sensitive adduct whose fluorescence increases as the environment becomes more hydrophobic. Addition of decarbamylated Rubisco complexed with ribulose 1,5-bisphosphate to the chemically modified mutant either in the presence or in the absence of ATP, ADP, or ATPγ had no effect on NBD fluorescence (data not shown). However, addition of low concentrations of ADP caused a marked quenching of the NBD fluorescence and a shift in the emission spectrum to a higher wavelength (Figure 4, traces a and b). The effect was saturated at low concentrations of ADP (i.e.